use S3 to upload tarball needed for DryRun

[belforte]

CRABServer/src/python/TaskWorker/Actions/DryRunUploader.py

Lines 45 to 55 in 411c812

	self.logger.info('Uploading dry run tarball to the user file cache')
	ufc = UserFileCache(mydict={'cert': kw['task']['user_proxy'], 'key': kw['task']['user_proxy'], 'endpoint': kw['task']['tm_cache_url']})
	result = ufc.uploadLog('dry-run-sandbox.tar.gz')
	os.remove('dry-run-sandbox.tar.gz')
	if 'hashkey' not in result:
	raise TaskWorkerException('Failed to upload dry-run-sandbox.tar.gz to the user file cache: ' + str(result))
	else:
	self.logger.info('Uploaded dry run tarball to the user file cache: ' + str(result))
	update = {'workflow': kw['task']['tm_taskname'], 'subresource': 'state', 'status': 'UPLOADED'}
	self.logger.debug('Updating task status: %s' % str(update))
	self.server.post(self.resturi, data=urllib.urlencode(update))

[belforte]

for this needs to introduce a new objecttype in S3 CRABCache

[belforte]

will start with current dry run sandbox even if too large and leave to a later time a more sensible optimization along the lines in #6544
But should start with a proper objecttype so we do not change again: #6552